three-factor capm risk exposures: some …web.usm.my/journal/aamjaf/vol 6-1-2010/6-1-3.pdf · some...
TRANSCRIPT
AAMJAF, Vol. 6, No. 1, 47–67, 2010
47
THREE-FACTOR CAPM RISK EXPOSURES:
SOME EVIDENCE FROM MALAYSIAN
COMMERCIAL BANKS
Aisyah Abdul Rahman
Faculty of Economics and Business,
Universiti Kebangsaan Malaysia,
43600, Bangi, Selangor
E-mail: [email protected]
ABSTRACT
This study investigates the determinants of the three-factor capital asset pricing model
(CAPM) risk exposures in the case of commercial banks. Five risk exposures are
examined namely, market, interest rate, exchange rate, total, and unsystematic risk
exposures. Our findings provide four major contributions. First, we find that different
types of risk exposures have different determinants. Second, the market risk exposure for
the Islamic bank in our study is lower than for conventional banks. Third, the merger
programme is fruitful because it reduces the interest rate risk exposure, total risk
exposure, and unsystematic risk exposure. Finally, our results show that the banks under
study have higher total and unsystematic risk exposures during the 1997 Asian financial
crisis. Thus, a clear understanding of this evidence helps in ensuring effective and
successful decision-making for regulators, policy makers and market players.
Keywords: banking, risk, merger, financial crisis
INTRODUCTION
Research found in banking literature has increasingly focused on risk exposure
since the 1997 Asian financial crisis. In fact, risk exposure has become one of the
most important factors in surviving the current, ongoing global economic crisis.
Although the scope of risk is vast, many studies focus on market, interest rate,
and foreign exchange rate risk exposure, as they are considered risks that can be
monitored, measured, and minimised with proper management. Among others,
one way to assess these risks is by using the three-factor capital asset pricing
model (CAPM), which evolved from the single-factor CAPM originally
developed by Sharpe (1964). Sharpe finds that the sensitivity to stock market
ASIAN ACADEMY of
MANAGEMENT JOURNAL
of ACCOUNTING
and FINANCE
Aisyah Abdul Rahman
48
return can determine asset pricing. However, Ross (1976) shows that stock
market return is not the only factor. Ross's arbitrage pricing theory (APT)
demonstrates that multiple factors significantly affect asset pricing. For the
banking sector, interest rate serves as the other established significant factor
(Stone, 1974; Martin & Keown, 1977; Brooth & Officer, 1985; Flannery &
James, 1984; Chance & Lane, 1980; Lynge & Zumwalt 1980; Jahankhani &
Lynge, 1980; Bae, 1990; and Lloyd & Shick, 1997). As financial globalisation
unfolds, a few studies have found that the exchange rate is another significant
factor (Chamberlain et al., 1997; Hahm, 2004; Francis & Hunter, 2004; Wong et
al., 2009; and Yong et al., 2009). Even though studies from developed countries
(Jianping & Saunders, 1995; Jianping & Ahyee, 1994; Allen et al., 1995; He et
al., 1996; and Lausberg, 2001) show that real estate investment return is another
factor, it is not significant for emerging countries, as real estate investment is still
at an early stage. Hence, market, interest rate, and exchange rate risk exposures
appear to be the significant factors in determining asset pricing for the Malaysian
banking sector.
It is important to take note that Malaysia applies a dual banking system
in which both conventional and Islamic banking products are offered to the same
customers. Hence, the customers are left with two choices in selecting which
banks they use to meet their banking needs. Their decisions may be influenced
by the risk exposures faced by the banks. To some extent, those risk exposures
depend on the background information of the banking systems. For the Islamic
banking industry, the first full-fledge Islamic bank was established in 1983,
known as Bank Islam Malaysia Bhd. (BIMB). It was given a monopoly status for
ten years to strengthen its position before the central bank allowed other Islamic
banks or windows to emerge in 1993. The Islamic banks are governed by the
Islamic Banking Act 1983 (IBA). As of January 2010, there are seventeen
Islamic banks, but only BIMB is listed in the Malaysia stock exchange.
Meanwhile, the conventional banking system began in Malaysia prior to its
independence and is governed by the Banking and Financial Institutions Act
1989 (BAFIA). In 1990, there were 22 local and 16 foreign banks altogether. In
order to encourage sustainability in the face of financial liberalisation, the
Malaysian central bank launched a consolidation programme in the late 1990s.
As of January 2010, there are 10 local and 13 foreign banks. Out of these, only
10 are listed in the stock exchange. As the estimation for CAPM risk requires
data on stock returns, this study uses data comprising all 11 bank-holding
companies listed in the stock exchange.
As burgeoning studies in this area centre on developed countries,
research on the Islamic banking industry is still relatively scarce. Most attention
is given either to conceptual frameworks or the macro-economic context. Many
studies simply presume that the behavioural attributes of Islamic banks are
Three-Factor CAPM Risk Exposures
49
similar to conventional banks. However, the fact that both banking systems have
different objectives, undergo different operational processes, and are bound by
different regulatory acts, it is unfair to generalise such findings. Indeed, Rahman
et al. (2009, 2008) show that the insolvency risk behaviour of Islamic banks is
different from conventional banks. Thus, the aim of this study is to investigate
the determinants of three-factor CAPM risk measures for Malaysian commercial
banks while at the same time acknowledging the different risk level of the Islamic
banking industry.
The contributions of this study are at least three-fold. Firstly, we evaluate
various risk exposures uniquely for the case of emerging markets like Malaysia.
Studying emerging markets is crucial, as there is no conclusive evidence
indicating the similarity of risk behaviour between emerging and developed
countries. On the one hand, Simpson (2002) finds that e-commerce risk,
efficiency, and rate of progress are significantly different between emerging and
developed countries. On the other hand, Joyce and Nabar (2009) show that global
economic crisis and financial liberalisation cause emerging economies to be
susceptible to the "sudden stops" on domestic investment due to the systemic
banking crisis from the developed countries, thereby suggesting similarities
between the two markets. In addition, as there are various types of bank risk
exposures (e.g., insolvency risk, subordinated debt implied risk, credit risk,
liquidity risk, and operational risk), we investigate the determinants of three-
factor CAPM risk namely, market risk, interest rate risk, foreign exchange rate
risk, unsystematic risk, and total risk exposure. Most prior CAPM research
examines the macro-economic factors affecting the asset pricing of the banking
industry, but it does not investigate the micro-economic features that may affect
the risk exposure of dynamic macro-economic factors. Finally, we also
investigate the different risk behaviour of the Islamic banking system.
The remainder of this paper is organised as follows: Section 2 outlines a
review of the literature on risk determinants, and section 3 highlights the data and
methodology. Section 4 reports the findings, and section 5 provides conclusion.
LITERATURE REVIEW
There are an enormous number of empirical studies on bank risk exposure.
Whereas Madura et al. (1994) and Ahmad and Ariff (2004) observe the
determinants of risk exposure per se, Saunders et al. (1990), Anderson and Fraser
(2000), Konishi and Yasuda (2004), Hassan (1993), Cebeyonan and Strahan
(2004), Gallo et al. (1996), Brewer et al. (1996), Gonzales (2004), Marco and
Fernandez (2008), Lepetit et al. (2008), and Yong, Faff and Chalmers (2009)
investigate various issues of bank risk exposure either in single-country or single-
Aisyah Abdul Rahman
50
region studies. There are some cross-country studies in this area, such as Dinger
(2009), Laeven and Levine (2009), and Angkinand and Wihlborg (2009).
Whereas single-country research includes bank-specific variables (BSV), cross-
country research often includes BSV and country-specific variables (CSV) as
control variables.
Madura et al. (1994) examine the determinants of the US-implied risk
exposure of deposit-taking institutions and commercial banks. Their findings for
depository institutions are not consistent with their findings for commercial banks.
With regards to depository institutions, real estate financing, real estate owned,
and level of capital buffer are the significant determinants. Meanwhile, for
commercial banks, the significant factors are real estate owned and capital buffer.
In another study, Ahmad & Ariff (2004) investigate the factors affecting the risk
exposure of Malaysian deposit-taking institutions using a single-factor CAPM.
Unlike Madura et al. (1994), they do not separately study commercial banks.
Their findings show that the determinants for market risk exposure include loan
default, cost of funds, loan expansion, and loan concentration. Meanwhile, the
determinants for unsystematic risk include loan default, cost of funds, and short-
term interest rate.
Studying how ownership structure affects US bank risk exposure,
Saunders et al. (1990) incorporate equity capital, fixed asset, and size as BSV.
They develop seven risk exposures based on a two-factor CAPM.1 Their findings
reveal that the BSV differently affect the seven types of risk exposure. One of
their interesting findings is that size is positively related to the market risk but
negatively related to the interest rate risk exposure. The underlying reason for
this conflicting result is that larger banks tend to be sensitive to market
movements, but at the same time, they are more able to diversify their interest
rate risk exposure.
Studying a similar issue, Anderson and Fraser (2000) estimate the single-
factor CAPM risk measure and insert an additional BSV, known as frequency.2
They argue that frequency represents the level of business risk exposure, as it
1 The risk measures include 1) total risk, 2) unsystematic risk for short-term interest rate, 3)
unsystematic risk for long-term interest rate, 4) market risk for short-term interest rate, 5) market
risk for long-term interest rate, 6) short-term interest rate, and 7) long-term interest rate risk
exposure.
The independent variables include 1) ownership structure (i.e., the proportion of stock held by
managers), 2) financial leverage [Capital/Total Asset (TA)], 3) operating leverage [Fixed
Asset/Total Asset (FA/TA)], and 4) size (TA). 2 The single-factor CAPM risk measures may be total risk, unsystematic risk, and systematic risk
exposures (i.e., total risk minus unsystematic risk exposures). Frequency is defined as the ratio of
an average daily share volume traded to number of shares outstanding.
Three-Factor CAPM Risk Exposures
51
denotes the speed at which new information is captured in the stock price and is
correlated to variances in bank balance-sheet and off-balance-sheet portfolios.
Their findings show that size is negatively related to total risk but positively
related to systematic risk; meanwhile, frequency is positively related to both total
and systematic risks.
Whereas Saunders et al. (1990) and Anderson and Fraser (2000) analyse
the case of US, Konishi and Yasuda (2004) and Marco and Fernandez (2008)
present comparable studies for the cases of Japan and Spanish, respectively.
Their studies add value to an existing gap in the literature with respect to risk
measurements.3 Whereas Konishi and Yasuda (2004) reveal that size and capital
buffer are significantly related to the two-factor CAPM and insolvency risk
exposures, Marco and Fernandez (2008) show that size, profitability, and
business types are significantly related to insolvency risk exposure. Regarding
cross-country research, Laeven and Levine (2009) and Angkinand and Wihlborg
(2009) find that size, credit quality, capital buffer, liquidity ratio, Gross
Domestic Product (GDP) growth, inflation, and interest rate are significantly
related to credit and insolvency risk exposures.
Examining the impact of mutual fund investments on US bank risk
exposure, Gallo et al. (1996) incorporate loan expansion, capital buffer, short-
term investment securities, and size as BSV. Estimating market and industry risk
exposures using the single-factor CAPM, they find that loan expansion is
significant for market and industry risk exposures, whereas short-term investment
securities are significant for industry risk exposure.
Looking into the impact of income structure on European banks risk
exposure, Lepetit et al. (2008) include size, capital buffer, profitability, loan
expansion, and cost of funds as BSV. Using the single-factor CAPM and
insolvency risk estimations, they find that size and capital buffer are significant
BSV.
Investigating how off-balance sheet activities affect bank risk, Hassan
(1993), Cebeyonan & Strahan (2004), and Yong et al. (2009) find that credit
expansion, credit quality, lending structure, cost of funds, Gap analysis, liquidity
ratio, capital buffer, and size are significant BSV. Hassan (1993) investigates the
3 Konishi and Yasuda (2004) analyse CAPM and insolvency risk whereas Marco and Fernandez
(2008) only focus on the insolvency risk exposure. The two-factor CAPM risk exposures include
1) total risk, 2) unsystematic risk for the short-term interest rate, 3) unsystematic risk for the
long-term interest rate, 4) market risk exposure for the short-term interest rate, 5) market risk
exposure for the long-term interest rate, 6) systematic risk exposure for the short-term interest
rate, and 7) systematic risk exposure for the long-term interest rate. The insolvency risk exposure
is the Z risk score.
Aisyah Abdul Rahman
52
impact of loan sales on US bank risk exposures based on the single-factor CAPM
and subordinated debt models.4 In contrast to Hassan (1993), Cebenoyan and
Strahan (2004) evaluate the loan sales-risk relationship based on financial ratios.5
Meanwhile, Yong et al. (2009) study the impact of derivative activities on risk
based on the three-factor CAPM approach.
DATA AND METHODOLOGY
This study adopts Generalised Least Squares (GLS) unbalanced panel regression
estimation. Three models are tested namely, pooled effect, fixed effect, and
random effect models. The best model is selected based on the Likelihood Ratio
and Hausman tests.6 To address heteroskedasticity, cross-section weighting is
used in the GLS estimation. Following Shahimi (2006) and Zakaria (2007), the
first-order autocorrelation problem is addressed based on the Park's model. The
data are comprised of all 11 listed commercial bank-holding companies in
Malaysia for 1994–2006. Table 1 displays the list of bank-holding companies, the
sampling period, and the completed merger period. The period from 2007 to 2009
are excluded due to the volatile market resulting from the food crisis, the oil price
crisis, and the US subprime crisis. The research design in this study is as follows:
it i it it it it itZ a I Y C (1)
Note that Z is the alternate measures of the three-factor CAPM risk exposures of
banks, X is a vector of BSV, I is a dummy for Islamic banking, Y are dummies for
the during and post-crisis periods, C is a dummy for the post-merger period, ai is
an individual-specific intercept, and β, γ, δ, and Φ are the slope coefficients to be
estimated.
The five alternate risk measures include market, interest rate, exchange
rate, total, and unsystematic risk exposures. (Detailed specifications of these risks
are discussed in the following subsection). We analyse the risk behaviour of the
Islamic banking industry using a dummy variable called DISLAM (Islamic bank
= 1; conventional banks = 0).7 We also take into account the financial crisis and
4 Please refer to Hassan (1993) for a detailed explanation of the implied asset subordinated debt
models. 5 The four risk measures are 1) ζROE,, 2) ζROA, 3) ζLLP./TL, and 4) ζnpl/TL. 6 Please refer to Beaver et al. (1989), Hsio (2002), Gujarati (2003), Shahimi (2006), and Zakaria
(2007) for further details on panel data regression techniques. 7 The Islamic bank's stock return is based on the stock prices of BIMB. As the majority of Islamic
banks are small and do not trade in Bursa Malaysia, BIMB is the only available sample
representing the Islamic banking industry in Malaysia. We cannot consider Islamic banking
Three-Factor CAPM Risk Exposures
53
consolidation effect during the period of study using the variables DCRISIS (year
1997 and 1998), DPOST-CRISIS (year 1999 to 2006), and DMERGER (post
merger = 1; pre-merger = 0). Most of the completed merger periods are either in
year 2000 or 2001, and thus, they do not overlap with the post-crisis period.
Table 1
List of Commercial Banks: Sampling Period and Completed Merger Period
Type of Bank Bank-Holding Company Sampling
Period
Completed
Merger Period1
Islamic bank Bank Islam Malaysia Bhd. (BIMB) 1994–2006 –
Conventional banks Hong Leong Bank Bhd. 1994–2006 2000
Malayan Banking Bhd. 1994–2006 2001
Public Bank Bhd. 1994–2006 2001
RHB Bhd. 2002
Affin Bank Bhd. 1994–2006 2000
Alliance Bank Bhd. 1994–2006 2001
Ambank Bhd. 1994–2006 2001
CIMB Bank Bhd. 1994–2006 2000
EON CAP Bhd. 1994–2006 2001
Southern Bank Bhd. 1994–2004 2001
Notes: 1Completed merger period is based on the final merger entity as announced in the annual reports of bank-
holding companies. See Said et al. (2008) for a detailed study of the evolution of banking groups and final entities after merger.
The selection of BSV in this study is closest in spirit to that of Hassan
(1993), Madura et al. (1994), Ahmad and Ariff (2004), Cebeyonan and Strahan
(2004), and Yong et al. (2009). (Appendix A presents a list of references that
have adopted similar variables).The BSV are TL, BPS, PLL, TE, GAP, INTEXP,
INV, LTA, NONII, and MGT. They are the best available proxies for loan
expansion, real estate lending, loan quality, capital buffer, GAP analysis, cost of
funds, liquid asset, size, non-interest income, and management efficiency,
respectively.8
windows and other Islamic subsidiaries of domestic bank-holding companies since they are not
listed in Bursa Malaysia, and hence, we do not have market prices. 8 Appendix A lists a number of studies that have employed similar BSV. We select the best
available variables for the case of Malaysia (that is, for a single-country study); thus,
insignificant variables and CSV are not included in the analysis. For the case of Islamic banks,
we let 1) interest expense be the sum of income distributed to depositor and shareholder fund and
2) non-interest income be the total income minus income from financing from both depositor and
shareholder funds.
Aisyah Abdul Rahman
54
The expected relationships among the BSV follow from previous studies.
For loan expansion, Hassan (1993), Madura et al. (1994), and Gallo et al. (1996)
suggest that the illiquidity of loan and the default risk exposure are the
underlying reasons for the positive relationship between TL and risk. For real
estate lending, Blasko and Sinkey Jr. (2006) show that real estate lending
increases the interest rate risk exposure, but Ahmad and Ariff (2004) find that
risky sector lending (the majority of which comes from the real estate sector)
reduces market risk exposure. Meanwhile, Rahman et al. (2008, 2009) show that
real estate lending reduces the insolvency risk exposure for Islamic banks, but the
reverse is true for conventional banks. Thus, the coefficient sign of BPS can be
negative or positive, depending on different types of risk exposures. With regards
to loan default, earlier studies hypothesise that PLL represents the probability of
future default and loan quality; thus, it is expected to be positively related to risk.
In the case of capital buffer, total equity is assumed to provide a cushion against
loss; hence, an increase in TE can reduce the bank risk exposure. For GAP
analysis, a positive GAP indicates that a particular bank is an asset-sensitive bank,
whereas a negative GAP indicates that it is a liability-sensitive bank. A positive-
GAP bank (that is, an asset-sensitive bank) is exposed to the risk that an interest
rate will fall, whereas a negative-GAP bank (that is, a liability-sensitive bank) is
exposed to the risk that an interest rate will increase. Thus, the greater the
absolute value of GAP is, the more the bank is exposed to changes in the interest
rate. Despite GAP analysis, Madura et al. (1994) argue that bank risk depends on
the proportion of funds obtained in the deposit account as measured by interest
expense, which is not captured in GAP analysis. They hypothesise that the higher
the deposit is, the higher the interest expense and the volatility of net interest
income are; thus, the bank is riskier. For the case of INV, it is linked to risk from
the perspective of deposit withdrawal. Maintaining idle cash is an opportunity
cost to banks, and thus, banks hold short-term investment securities to standby
the need for extraordinary deposit withdrawal, suggesting a negative association
between INV and risk. With regards to size, Saunders et al. (1990) and Hassan
(1993) argue that the greater the size, the greater the potential to diversify
business risk, thereby reducing bank risk exposure. However, Anderson and
Fraser (2000) suggest that the impact of size on risk depends on the lending
structure. If the loan composition is the same, larger banks should have lower risk
as compared to smaller banks. Nonetheless, if the loan structure is different, a
larger bank may have higher risk exposure than a smaller one due to its tendency
to embark into riskier lending sectors that can yield higher returns. Similarly,
Gonzales (2004) mentions that with the existence of the economy of scale,
increased market power, and "too big to fail" policies for big banks, a larger bank
tends to enter into risky activities either through lending strategies or off-balance
sheet activities. Against this background, size can be positively or negatively
related to risk exposure. For the case of non-financing activities, Madura et al.
(1994) offer evidence that diversifying from the traditional role banking (that is,
Three-Factor CAPM Risk Exposures
55
providing loans) reduces bank risk exposure. With regards to management
efficiency, Angbazo (1997) and Ahmad and Ariff (2003, 2004) believe that the
increasing efficiency of management reduces bank risk exposure. With regards to
Islamic banking, even though most of the Islamic banking research conceptually
highlights the additional risk exposure faced by the Islamic bank, this heightened
risk exposure has not yet been empirically tested. Considering this, DISLAM can
be positive or negative, depending on the development stage. For the crisis
dummies, risk exposures are expected to be higher during the crisis period but
may be higher or lower during the post-crisis period, depending on the recovery
process. Finally, if the benefit of bank consolidation indeed materialised,
DMERGER is expected to be negative.
SPECIFICATION FOR THREE-FACTOR CAPM RISK EXPOSURES
Our three-factor CAPM risk measure is akin to that employed by Hahm (2004),
Francis & Hunter (2004), Wong et al. (2009), and Yong et al. (2009). It is
expressed as follows:
Rt = α + βm (Rmt) + βi (Rit) + βforex (Rforex) + εt (2)
Note that:
Rt = daily stock return of bank-holding companies during a one-year period.
Rmt = daily Kuala Lumpur Composite Index (KLCI) market return during a
one-year period.
Rit = daily Malaysian Government Securities (MGS) 10-year return during a
one-year period.9
Rforext = daily nominal effective exchange rate (NEER) return during a one-year
period.
et = the error term that captures all other factors that affect bank return that
are not taken into account explicitly.
αj = the intercept of the characteristic line.
From the above equation, 5 yearly risk measures from 1994 to 2006 are estimated
for 11 Malaysian bank-holding companies. The five risk exposures are:
9 Daily return on 10-year MGS represents the long-term interest rate risk exposure during a one-
year period. We also analyse the short-term interest rate risk exposure using the overnight return
of Kuala Lumpur Interbank Offer Rate (KLIBOR), but our finding fails the F-statistic test. We
suspect that is due to very insignificant daily changes in the KLIBOR overnight rate. The results
for short-term interest rate risk exposure are provided upon request.
Aisyah Abdul Rahman
56
(i) Market risk exposure (βm)
(ii) Long-term interest rate risk exposure (βi)
(iii) Exchange rate risk exposure (βforex)
(iv) Unsystematic risk exposure (standard deviation of εt )
(v) Total risk exposure (standard deviation of Rt)
After the five risk measures are estimated, we then analyse our research design
based on Equation 1 using the GLS unbalanced panel regression estimation with
yearly data.
FINDINGS
The descriptive statistics as well as the correlation matrix are displayed in Tables
2 and 3, respectively. As the correlation matrix in Table 3 shows lower values
than 0.8, we consider that all independent variables are not seriously correlated;
hence, all variables are regressed simultaneously.10
For the regression results, the
fixed effect model appears to be the best model for all types of CAPM risk
exposures.11
Hence, the discussion in this section is based on the fixed effect
model. Table 4 presents the fixed effect estimations for market, long-term interest
rate, exchange rate, total, and unsystematic risk exposures.
For the case of market risk exposure, Table 4 shows that total loan
expansion (TL) is the only significant factor. The positive relationship of TL is as
expected and consistent with Hassan (1993), Gallo et al. (1996), Ahmad and
Ariff (2003, 2004), and Gonzales (2004). For DISLAM, our results show that the
Islamic banking system has lower risk exposure than the conventional banking
system. A lower market risk exposure for the Islamic banking industry may be
due to heavy reliance on the fixed profit rate system. Even though nowadays
there is another product innovation that offers a quite similar function to a floated
profit rate system (that is, the principal of Musyarakah Mutanakisah), its use in
the market is still new and limited, and thus, its influence on market risk exposure
is insignificant.
10 Gujarati (2003) sets a cut of point of 0.8 for the correlation coefficient matrix. Values higher than
0.8 indicate that the variables are strongly correlated; thus, they should be analysed to address
issues of multicollinearity. 11 The results for the no effect model, random effect model, the Likelihood ratio test, and the
Hausman test are provided upon request.
Table 2
Descriptive Statistics
Variables Mnemonics Mean Std. Dev
Risk Indicator (dependent variable)
Market Risk Exposure KLCI 1.077 0.404
Interest Rate Risk Exposure MGS 0.006 0.114
Exchange Rate Risk Exposure NEER 0.013 0.561
Total Risk Exposure TRISK 0.025 0.013
Unsystematic Risk Exposure UNSYS 0.019 0.008
Credit-related Variables
Ratio of Total Loans to Total Asset TL 0.520 0.250
Ratio of Broad Property Sector to Total Loan BPS 0.393 0.082
Ratio of Provision of Loan Loss to Total Asset PLL 0.009 0.010
Capital-related Variables
Ratio of Total Equity to Total Asset TE 0.085 0.037
Interest Rate-related Variables
Ratio of GAP to Total asset GAP –0.166 0.187
Ratio of Interest Expense to Total asset INTEXP 0.031 0.014
Liquidity-related Variable
Ratio of Short Term Investment to Total Asset INV 0.140 0.122
Business Operation-related Variables
Log of Total Asset LTA 7.318 0.440
Ratio of Non-interest Income to Total Asset NONI 0.009 0.006
Ratio of Earning Asset to Total Asset MGT 0.852 0.101
Table 3
Correlation Matrix of Independent Variables
(continued)
TL BPS PLL TE GAP INTEXP INV LTA NONII MGT
TL 1
BPS –0.040 1
PLL 0.502 –0.085 1
TE –0.076 0.426 –0.122 1
GAP 0.019 0.029 0.126 0.013 1
INTEXP 0.321 –0.154 0.540 –0.146 0.232 1
INV 0.309 0.022 0.231 –0.270 –0.255 0.230 1
LTA –0.276 –0.295 –0.185 –0.172 –0.322 –0.442 –0.054 1
Aisyah Abdul Rahman
58
Table 3 (continued)
For the case of long-term interest rate risk exposure, INTEXP and MGT show
significant effects. Consistent with prior studies, INTEXP is positively related to
interest rate risk. If a bank has a large amount of interest expenses, it is exposed
to greater interest rate movements, thus increasing its risk exposure. MGT is
negatively related to management efficiency, suggesting that banks are efficient
in managing their long-term interest rate risk exposure. For DISLAM, our result
shows that the risk behaviour of the Islamic banking industry is different from the
conventional banks. To the extent that Islamic banks benchmark their profit rate
to the market interest rate, their interest rate risk behaviour will always be similar
to that of conventional banks.
Table 4
The GLS Fixed Effect Estimation for the CAPM Risk Exposures
Expected
coefficient sign
Market
risk exposure
Interest
rate risk exposure5
Exchange
rate risk exposure
Total risk
exposure
Unsystematic
risk exposure
C
–1.772
(–1.069)
0.332
(0.876)
–2.869
(–0.415)
3.424***
(2.862)
2.362**
(2.088)
TL
+ 0.291*
(1.698)
0.044
(0.335)
–0.797
(–0.767)
0.001
(0.006)
–0.023
(–0.120)
BPS
+/– –0.045
(–0.294)
–0.174
(–0.992)
0.117
(0.120)
–0.035
(–0.166)
0.050
(0.263)
PLL
+ –0.032
(–0.721)
–2.170
(–1.658)
0.851**
(–2.258)
0.0178
(0.351)
0.019
(0.382)
TE
– –0.035
(–0.763)
0.254
(0.814)
–0.265
(–0.870)
–0.109**
(–2.605)
–0.097**
(–2.244)
GAP
+ –0.018
(–0.606)
–0.011
(–0.122)
–0.362**
(–2.162)
0.026
(1.042)
0.021
(0.982)
INTEXP
+ –0.082
(–0.728
8.186***
(3.158)
0.728
(1.363)
0.117
(1.284)
0.110
(1.282)
INV
– –0.047
(–1.415)
0.218
(1.121)
–0.278
(–1.384)
–0.081**
(–1.999)
–0.064
(–1.559)
TA
+/– 0.129
(0.539)
1.035
(0.602)
–1.201
(–1.512)
–1.119***
(–6.239)
–0.946***
(–5.551)
(continued)
TL BPS PLL TE GAP INTEXP INV LTA NONII MGT
NONII –0.342 –0.056 –0.309 0.071 0.133 –0.182 –0.260 –0.125 1
MGT 0.389 –0.017 0.189 –0.052 –0.184 0.249 0.214 –0.456 0.175 1
Three-Factor CAPM Risk Exposures
59
Table 4 (continued)
Expected
coefficient sign
Market
risk exposure
Interest
rate risk exposure5
Exchange
rate risk exposure
Total risk
exposure
Unsystematic
risk exposure
NONII
– 0.044
(0.738)
0.003
(0.126)
–0.849*
(–1.795
–0.078
(–1.420)
–0.062
(–1.310)
MGT
+/– –0.050
(–0.079)
–0.622**
(–2.143)
–1.623
(–0.392)
0.838
(1.598)
0.906*
(1.877)
DISLAM
+/– –0.372**
(–2.156)
–0.015
(–0.268)
–0.206
(–0.299)
–0.163
(–1.504)
–0.075
(–0.750)
DMERGER
– 0.033
(0.686)
–0.058*
(–1.751)
–0.184
(0.196)
–0.004*
(–1.889)
–0.004***
(–2.987)
DCRISIS
+ 0.211
(1.433)
–0.176
(0.404)
0.674***
(6.436)
0.477***
(4.730)
DPOST-CRISIS
+/– 0.113
(0.760)
–0.148
(0.476)
0.435***
(3.796)
0.330***
(3.152)
Adj R-squared 0.560 0.194 0.137 0.794 0.760
Prob (F-Stats) 0.000 0.021 0.029 0.000 0.000
D.W. statistics 1.980 1.832 2.603 1.911 1.978
Notes:
1. The dependent variables are market, long-term interest rate, exchange rate, total, and unsystematic risk exposures.
2. Figures in parentheses are t-statistics. 3. ***, **, and * denotes significance at 1%, 5%, and 10% confidence levels, respectively.
4. The findings for the GLS pool effect and GLS random effect models are available upon request.
With regards to exchange rate risk exposure, our result indicates that PLL, GAP,
and NONII are significant factors. The positive relationship of PLL implies that
as a bank increases its provision for loan losses, its exchange rate risk exposure
increases, suggesting that the Malaysian banks are vulnerable to external shock.
Note that increasing PLL implies decreasing loan quality. Decreasing loan
quality can be due to foreign borrowers who obtain loans denominated in foreign
currency or from local borrowers whose businesses are affected by the exchange
rate movement. The inverse relationship of GAP reveals that an increase in the
maturity mismatch of net short-term assets or liabilities is associated with a
decrease in exchange rate risk exposure. As GAP shows the net absolute position
of short-term assets and liabilities, a positive-GAP bank (that is, an asset-
sensitive bank) is exposed to the risk that the interest rate will fall, whereas a
negative-GAP bank (or a liability-sensitive bank) is exposed to the risk that the
interest rate will increase.12
During the period of study, the increasing
12
GAP = [net fed funds sold + trading account securities + securities maturing in less than one year
+ loan and leases maturing in less than one year + customer liabilities to the bank for outstanding
Aisyah Abdul Rahman
60
dependency of net short-term position on foreign denominated assets or liabilities
is beneficial, as it reduces exchange rate risk exposure. However, it should be
noted that risk indirectly depends on the accuracy in forecasting foreign-
denominated asset pricing; thus, banks should be aware of global economic
developments. Finally, the inverse relationship of NONII implies that the
increasing involvement of banks in fee-based activities reduces exchange rate
risk exposure, as most of the off-balance-sheet activities are self-liquidating
contingencies.
Regarding total risk exposure, TE, INV, and LTA are significant
variables. The negative relationships of TE and INV are as expected and
consistent to prior studies. Nonetheless, the inverse relationship of LTA implies
that as a bank increases its size, it reduces its total risk exposure. This finding
contradicts the results given by Saunders et al. (1990), Hassan (1993), and
Gonzalez (2004), but it is consistent with Anderson and Fraser (2000) and
Konishi and Yasuda (2004). Recall that total risk is a summation of systematic
(that is, market, interest rate, and exchange rate risk) and unsystematic risk. As
our findings show that LTA is not significant for systematic risk exposure but is
significant for unsystematic risk exposure. This implies that the impact of size on
the total risk exposure is greatly influenced by unsystematic risk exposure. This
inverse relationship implies that a larger Malaysian bank is more capable of
diversifying its unsystematic risk, which then reduces its total risk exposure.
For unsystematic risk exposure, the significant factors are TE, LTA, and
MGT. These variables are similar to total risk exposure, except for MGT. As total
risk exposure comprises both systematic (i.e. market, interest rate, and exchange
rate risk) and unsystematic risk, it can be inferred that the positive relationship
between MGT and unsystematic risk exposure is offset by the negative
relationship between MGT and interest rate risk exposure, which produces an
insignificant relationship between MGT and total risk exposure.
CONCLUSION
This study offers four contributions. First, as different types of risk exposure have
different risk determinants, market players should prioritise their risk
management based on their mission and vision. If a bank takes aggressive
strategies aimed at international activities, its main concern should be how to
efficiently manage the determinants of exchange rate risk exposure rather than
acceptances] – [domestic and foreign deposit less than one year + CDs less than one year + other
borrowed money + the bank liabilities on customer acceptances outstanding]. The absolute value
of GAP is deflated by total assets.
Three-Factor CAPM Risk Exposures
61
other types of risk exposure. Second, even though our findings show that the risk
behaviours of Islamic and conventional banking systems are similar for all types
of risk exposure except market risk exposure, it is worth mentioning that in the
upcoming years, the Islamic banking industry may behave differently as grows in
number and size. With the assumption that Islamic banks would eventually based
on practices different from those of conventional banks (particularly in
determining the profit rate return), we believe that the determinants of the interest
rate, total risk, and unsystematic risk exposures will differ greatly in the future.
Third, our findings show that bank mergers seem to benefit market players, as
they reduce interest rate, total, and unsystematic risk exposures. Finally, our
findings provide empirical evidence suggesting that the 1997 financial crisis
increased total and unsystematic risk exposures. With this in mind, policy makers
as well as market players in the banking industry should be better prepared to
face the ongoing global financial crisis by proactively introducing prudent risk
management guidelines to prevent the Malaysian banking system from repeating
the US subprime crisis.
REFERENCES
Ahmad, N. H., & Ahmad, S. N. (2004). Key factors influencing credit risk of Islamic
bank: A Malaysian case. Journal of Muamalat and Islamic Finance Research,
1(1), 65–80.
Ahmad, N. H., & Ariff, M. (2004). Key risk determinants of listed deposit-taking
institutions in Malaysia. Malaysian Management Journal, 8(1), 69–81.
———. (2003). What factors determine the total risk of deposit-taking institutions? Paper
presented in MFA's 5th Annual Symposium, 23–24 April, 888–902. Cyberjaya:
MMU.
Allen, M., Madura, T., & Jeff, W. K. J. (1995). Commercial banks exposure and
sensitivity to the real estate market. Journal of Real Estate Research, 10(2),
129–140.
Anderson, R. D., & Fraser, D. R. (2000). Corporate control, bank risk taking, and the
health of the banking industry. Journal of Banking and Finance, 24(8), 1383–
1398.
Angbazo, L. (1997). Commercial bank net interest margin, default risk, interest rate risk,
and off-balance sheet banking. Journal of Banking and Finance, 21(1), 55–87.
Angkinand, A., & Wihlborg, C. (2009). Deposit insurance coverage, ownship, and banks'
risk-taking in emerging markets. Journal of International Money and Finance,
29(2), 1–23. doi:10.1016/j.jimonfin.2009.08.001
Bae, S. C. (1990). Interest rate changes and common stock returns of financial institutions:
Revisited. The Journal of Financial Research, XIII(1), 71–79.
Beaver, W., Eger, c., Ryan, S., and Wolfson, M. (1989). Financial Reporting.
Supplemental disclosures and bank share prices. Journal of Accounting Research,
27, 15–178.
Aisyah Abdul Rahman
62
Blasko, M., & Sinkey Jr. J. F. (in press) Bank asset structure, real-estate lending, and risk
taking. The Quartely Review of Economic and Finance.
Brewer, E., William, E. J., & Thomas, S. M. (1996). Risk, regulation, and S & L
diversification into non-traditional assets. Journal of Banking and Finance,
20(4), 723–744.
Brewer, E. (1989). Relationship between bank holding company risk and non-bank
activity. Journal of Economics and Business, 41, 337–353.
Brooth, J. R., & Officer, D. T. (1985). Expectations, interest rates, and commercial bank
stocks. The Journal of Financial Research, XIII(1), 51–58.
Cebeyonan, A. S., & Strahan, P. E. (2004). Risk management, capital structure and
lending at banks. Journal of Banking and Finance, 28, 19–43.
Chamberlain, S., Howe, J., & Popper, H. (1997). The exchange rate exposure of US and
Japanese banking institutions. Journal of Banking and Finance, 21, 871–892.
Chance, D. M., & Lane, W. R. (1980). A re-examination of interest rate sensitivity in the
common stocks of financial institutions. Journal of Financial Research, 3(1),
49–56.
Dinger, V. (2009). Do foreign-owned banks effect banking system liquidity risk? Journal
of Comparative Economics, 41(2&3), doi: 10.1016/j.jce.2009.04.03
Flannery, M. J., & James, C. M. (1984). The effect of interest rate changes on the
common stock returns of financial institutions. The Journal of Finance, XXXIX
(4), 1141–1153.
Francis, B. B., & Hunter, D. M. (2004). The impact of the Europe on risk exposure of the
world's major banking industries. Journal of International Money and Finance,
23, 1011–1042.
Gallo, J. G., Apilado, V. P., & Kolari, J. W. (1996). Commercial bank mutual fund
activities: Implications for bank risk and profitability. Journal of Banking and
Finance, 20(5), 1775–1791.
Gonzales, F. (2004). Bank regulation and risk-taking incentives: An international
comparison of bank risk. Journal of Banking and Finance, 29, 1153–1184.
Gujarati, D. N. (2003). Basic Econometrics (4th ed.). Singapore: McGraw-Hill Higher
Education.
Hahm, J. H. (2004). Interest rate and exchange rate exposures of banking institutions in
pre-crisis Korea. Applied Economic, 36(13), 1409–1419.
Hassan, M. K. (1993). Capital market tests of risk exposure of loan sales activities of
large US commercial banks. Quarterly Journal of Business and Economics, 27–
49.
He, L. T., Myer, F. C. N., & Webb, J. R. (1996). The sensitivity of banks stock returns to
real estate. Journal of Real Estate Finance and Economics, 12(2), 203–220.
Hsio, C. (2002). Analysis of Panel Data (2nd ed.). NY: Cambridge University Press
Jahankhani, A., & Lynge, M. J. (1980). Commercial bank financial policies and their
impact on market-determined measures of risk. Journal of Financial Research,
11(3),169–78
Jianping, M., & Saunders, A. (1995), Bank risk and real estate: An asset pricing
perspective. Journal of Real Estate Finance and Economics, 10(3), 199–224.
Jianping, M., & Ahyee, L. (1994). Is there a real estate factor premium? Journal of Real
Estate Finance and Economics, 9(2), 113–126.
Three-Factor CAPM Risk Exposures
63
Joyce, J. P., & Nabar, M. (2009). Sudden stops, banking crises and investment collapses
in emerging market. Journal of Development Economics, 90(2), 314–322
Konishi, M., & Yasuda, Y. (2004). Factors affecting bank risk taking: Evidence from
Japan. Journal of Banking and Finance, 28(1), 215–232
Lausberg, C. (2001). The real estate market risk of banks: Evidence of its importance and
consequences for managing risk in real estate lending. Article submitted to the
2001 annual meeting of the European Financial Management Association.
Working paper series in SSRN.
Laeven, L., & Levine, R. (2009). Bank governance, regulation and risk-taking. Journal of
Financial economics, 93(2), 259–275.
Lepetit, L., Nys, E., Rous, P., & Tarazi, A. (2008). Bank income structure and risk: An
empirical analysis of European banks. Journal of Banking and Finance, 32(8),
1452–1467.
Lloyd, W. P., & Shick, R. A. (1997). A test of Stone's two-index model of returns.
Journal of Financial and Quantitative Analysis, 12, 731–741.
Lynge, M., & Zumwalt, K. (1980). An empirical study of the interest rate sensitivity of
commercial banks returns: A multi-factor approach. Journal of Financial
Research and Qualitative Analysis, 15, 731–742.
Madura, J., Martin, A. D., & Taylor, D. A. (1994). Determinants of implied risk of
depository institutions. Applied Financial Economics, 4(5), 363–370.
Marco, T. G., & Fernandez, M. D. R. (2008). Risk-taking behaviour and ownership in the
banking industry: The Spanish evidence. Journal of Economics and Business, 60,
332–354.
Martin, J. D., & Keown, A. J. (1977). Interest rate sensitivity and portfolio risk. Journal
of Financial and Quantitative Analysis, 12(6), 181–191.
Rahman, A., Ibrahim, M., & Meera, A. K. M. (2009). Lending structure and bank
insolvency risk: A comparative study between the Islamic and Conventional
Banks. Journal of Business & Policy Research, 4(2), 189–211.
———. (2008). Does the Islamic Bank lending structure matter? Paper presented in 3rd
Islamic banking, Accounting & finance (iBAF) Conference 2008. Financial
Intelligence in Wealth Management: Islam Hadhari's Perspective, 29–30th July
2008, The Legend Hotel, Kuala Lumpur.
Ross, S. (1976). The arbitrage theory of capital asset pricing. Journal of Economic Theory,
13(3), 341–60.
Said, M., Rasidah, M., Fauzias, N., Low, S. W., & A. Rahman, A. (2008). The efficiency
effects of mergers and acquisitions in Malaysian Banking Institutions. Asian
Journal of Business and Accounting, 1(1), 47–66.
Saunders, A., Strock, E., & Travlos, N. G. (1990). Ownership structure, deregulation and
bank risk taking. Journal of Finance, 45(2), 643–654.
Shahimi, S. (2006). Penentu margin pendapatan bersih Bank Islam. PhD Dissertation,
Universiti Kebangsaan Malaysia, 143.
Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium under
conditions of risk. Journal of Finance, 19(3), 425–442.
Simpson, J. (2002). The impact of the Internet in banking: Observations and evidence
from developed and emerging markets. Telematics and Informatics, 19(4), 315–
330
Aisyah Abdul Rahman
64
Stone, B. K. (1974). Systematic interest rate risk in a two-index model of returns. Journal
of Financial and Quantitative Analysis, 9(5), 709–721.
Wong, T. C., Wong, H. J., & Leung, P. (2009). The foreign exchange exposure of
Chinese Bank. China Economic Review, 20(2), 174–185.
Yong, H. H. A., Faff, R., & Chalmers, K. (2009). Derivative activities and Asia-Pacific
banks' interest rate and exchange rate exposures. Journal of International
Financial Markets, Institutions, and Money, 19(1), 16–32.
Zakaria, R. H. (2007). Banks' securitization involvement, bank lending and bank stability:
A panel data analysis. PhD Dissertation, Universiti Kebangsaan Malaysia, 130.
APPENDIX A
Independent Variables Adopted In Bank Risk Research
Author(s) and Publication Date Dependent Variables Independent Variables
1 Ahmad, N. H., & Ariff, M.
(2004). Key risk determinants
of listed deposit-taking
institutions in Malaysia.
Malaysian Management
Journal, 8(1), 6–81.
4 risk measures:
β = market risk
standard deviation of the
error term of the regression
equation = unsystematic
risk
standard deviation of bank
return = total risk
Ratio of Book Value to
Market Value of equity =
equity risk.
NPL/TL
MGT: as in Angbazo
LEV: Tier2/Total Capital
(data started in year 2002)
RISKY= (BPS + Purchase
of securities loan +
consumption credit loan)
Regulatory Cap: Tier 1/TL
(data started in year 2002)
Interest expense
PLL: Provision for Loan
Loss
RWA: Risk Weighted Asset
IBOR: Interbank rate
SPREAD: MGS rate- T-Bill
GAP: (RSA-RSL)/TA
Loan expansion:
Loan/Deposit:
Size: Log of TA
2 Madura, J., Martin, A. D., &
Taylor, D. A. (1994).
Determinants of implied risk of
depository institutions. Applied
Financial Economics, 4, 363–
370
The ex-ante risk: using
average daily implied
standard deviation (ISD) of
the jth bank in year t.
(followed Latane &
Rendleman (1976)
ISD = based on call option
price.
Disadvantage of ISD: since
it is implied, it may be
different to the actual risk of
the firm.
Loan Expansion: TL/TA
Real Estate Financing: RE
loan/TA
RE owned/TA (no data in
Malaysia)
Capital buffer: TE/TA
PLL: PLL/TA
GAP =(RSA/RSL)
Interest expense/TA
Non-interest income/TA
3 Saunders, A., Strock, E., &
Travlos, N. G. (1990).
Ownership structure,
deregulation and bank risk
taking. Journal of Finance.
45(2) June 643–654
7 risk measures:
o σs = tot risk
o σes =unsys for s/t
o σel= unsys for l/t
o βms = sys for s/t
o βml= sys for l/t
Ownership structure -
proportion of stock held by
managers – expected to be
(+) related to risk.
Capital Buffer: TE/TA
Operating Leverage –
o βis=sys for s/t
o βil= sys for l/t
FA/TA (in significant data
in Malaysia)
Size –log of TA
4 Hassan, M. Kabir. (1993).
Capital market tests of risk
exposure of loan sales activities
of large US commercial banks.
Quarterly Journal of Business
and Economics, 27–49
β= systematic risk
σ= standard deviation of
equity return
DRM (default Risk
Premium) of subordinated
debt
Bank Implied Asset Risk
(Ronn-Verma Option
Pricing Model)
Bank Implied asset Risk
(Gorton-Santomero debt
pricing method)
Loan sales: Loan sale/TA
(no data in Malaysia)
Capital buffer: TE/TA
Specialisation index/TA
Loan Quality: PLL/TA
GAP: (RSA-RSL)/TA
Size: Log of TA
Dividend Payout Ratio/TA
(no data in Malaysia)
5 Gallo, John G., Apilado,
Vincent P., & Kolari, James W.
(1996). Commercial bank
mutual fund activities:
Implications for bank risk and
profitability. Journal of Banking
and Finance, 20, 1775–1791
3 risk measurements:
β(market risk: Wilshire
5000 Index, systematic risk)
FINL(industry risk:
Wilshire finance Index,
systematic risk)
Standard deviation of the
error term of the regression
equation (unsystematic risk)
Investment securities/TA
Loan Expansion: TL/TA
(Sales Fed-purchased
Fed)/TA (insignificant in
Malaysia)
Capital buffer: TE/TA
Size: log of TA
Profitability: interest
income/TA
NONII: NONII/TA
FEES – Mutual fund
Fee/TA
6 Angbazo, Lazarus. (1997).
Commercial bank net interest
margin, default risk, interest
rate risk, and off-balance sheet
banking. Journal of Banking
and Finance, 21, 55–87.
NIM= Net Interest
Income/average earning
asset.
NII= interest income –
interest expense.
Loan Quality: NCO/TL (in
consistent NCO data in
Malaysia)
Liquid risk: Investment
securities/TL
Capital Buffer: (Tier 1+
Tier 2)/TA (data started in
year 2002)
Cost of funds: (non-interest
expense- non-interest
income) /earning asset
Non-interest bearing
reserve/TA (no data in
Malaysia)
Mgt efficiency: Earning
asset/TA
7 Anderson, Ronald D., & Fraser,
Donald R. (2000). Corporate
control, bank risk taking, and
the health of the banking
industry. Journal of Banking
and Finance, 24(8), 1383–1398
3 risk measures:
total risk (σSP)
unsystematic risk (σε)
systematic risk (total risk –
unsystematic risk)
Ownership variable (% of
shares held by unaffiliated
blockholders)
Outside-blockholders
Size: log of TA
Frequency: average daily
share volume
traded/number of shares
outstanding. (market data)
Tobin Q:
Σ(CSmv+Liabilitybv/TA)
(market data)
8 Cebeyonan, A. Sinan &
Strahan, Philip E. (2004). Risk
management, capital structure
and lending at banks. Journal of
Banking and Finance, 28, 19–
43
Use four financial ratio
measures:
σROE
σROA
σLLP./TL
σnpl/TL
Loan sales (no data and
insignificant in Malaysia)
Loan purchase (no data and
insignificant in Malaysia)
Capital buffer: TE/ Earning
asset
Liquid Asset: Investment
securities/TA
Commercial + industrial
loan) / TA
Real estate loan /TA
9 Gonzales, Francisco. (2004).
Bank regulation and risk-taking
incentives: An international
comparison of bank risk.
Journal of Banking and
Finance, 29, 1153–1184.
2 proxies:
total risk exposure:
σStockReturn
credit risk exposure:
NPL/TL
Size: log of TA
FA/TA (insignificant in
Malaysia)
Investment in
unconsolidated subsidiaries
(insignificant in Malaysia)
Liability: Total Debt/TA
Tobin's Q: (market data)
REG – annual index of
freedom in finance &
banking activities (no data
in Malasyia)
LAW and Order: annual
index (no data in Malaysia)
Legal origin (insignificant
in Malaysia)
10 Konishi, Masaru & Yasuda,
Yukihiro. (2004). Factors
affecting bank risk taking:
Evidence from Japan. Journal
of Banking and Finance. 28,
215–232
Risk measures:
total risk: (σSR)
unsystematic risk: (σerror )
systematic risk: (total risk -
unsystematic risk)
market risk: (β1)
interest rate risk: (β2)
Insolvency risk: (Z-score
developed by Boyd et al.
(1993)- using market data
Size – log of TA
Frequency: volume o
shares/number of shares
outstanding
Akumudari officers –
retired high rank ministry of
finance and bank of Japan
(not applicable in Malaysia)
Franchise value: (market
data)
Capital constraint: (dummy)
11 Marco, Teresa Garcia and
Fernandez, M. Dolores Robles.
(2008). Risk-taking behaviour
and ownership in the banking
industry: The Spanish evidence.
Journal of Economics and
Business, 60, 332–354
Z risk index Ownership structure:
(dummy)
Diversification: Herfindahl
index
Public control: (dummy)
Turnover of government
structure: (dummy)
Profitability: ROE
Loan Expansion: TL/TA
Size: Log of TA
12 Dinger, Valeriya. (2009) Do
foreign-owned banks affect
banking system liquidity risk?
Journal of Comparative
Economics, 41(2&3)
doi:10.1016/j.jce.2009.04.003
Liquidity risk Transnational: (dummy)
Size: log of TA
Capital buffer: TE/TA
Foreign Asset/TA
(insignificant in Malaysia)
Loan Expansion: TL/TA
Fiscal policy (cross-country
study)
GDP growth (cross-country
study)
Per capita GDP
(cross-country study)
13 Angkinand, Apanard, &
Wihlborg, Class. (2009)
Deposit insurance coverage,
ownership, and banks’ risk-
taking in emerging markets.
Journal of International Money
and Finance, 1–23.
Doi:10.1016/j.jimonfin.2009.08
.001
NPL/CAP
Std dev of NPL/CAP
Z-score
Capital buffer: TE/TA
Real GDP/Cap (cross-
country study)
Real GDP growth (cross-
country study)
M2/Reserve (cross-country
study)
Inflation(cross-country
study)
Real interest rate
(cross-country study)
14 Yong, Hue Hwa Au, Faff,
Robert., & Chalmers, Keryn.
(2009). Derivative activities and
Asia-Pacific banks' interest rate
and exchange rate exposures.
International Financial
Markets, Institutions and
Money, 19,
16–32.
Interest rate risk
Exchange rate risk
TDER – derivative/TA
(insignificant in Malaysia)
Size: log of TA
Capital buffer: TE/TA
liquid asset: Investment
securities/TA
Profitability: net interest
income/TA
NONII: NONII/Total
income
Loan expansion: TL/TA
PLL: PLL/TA
15 Laeven, Luc., & Levine, Ross.
(2009). Bank governance,
regulation and risk-taking.
Journal of Financial
economics, 93, 259–275
Z-score
Std deviation of ROA
Equity volatility
Earning volatility
Tobin's Q (market data)
Profitability: Revenue
growth
Per capita income (cross-
countries study)
Rights (no data in
Malaysia)
Capital buffer: TE/TA
Capital stringency (no data
in Malaysia)
Restrict (no data in
Malaysia)
DI: (dummy for insurance)